SERVICE 08 / NUMBERS PEOPLE TRUST ENOUGH TO ACT ON
One version of the numbers
Reporting only helps when people believe it. We fix the definitions and the plumbing first, then build dashboards worth opening.
THE BUSINESS CHALLENGE
What this work
is for.
Two reports go into the same meeting with different revenue figures, and the meeting becomes about the discrepancy. Everyone leaves without deciding anything, and someone spends the next day on reconciliation.
WHAT WE DO
Reliable reporting is less about charts than about agreement. We establish where each number comes from, write down the definitions, and build a semantic layer so that revenue means one thing wherever it is used. On top of that we build dashboards aimed at specific decisions rather than general curiosity, and we show data freshness on the page so nobody acts on yesterday's figures believing they are today's. Where the underlying data cannot support a question yet, we say so instead of producing a confident-looking chart.
BUSINESS PROBLEMS THIS ANSWERS
- Reports that disagree. Two teams calculate the same metric differently and both are defensible. Trust in all reporting drops, not just the two.
- No agreed source of truth. Figures are assembled from exports, each refreshed on a different day. Nobody can say which system should win a conflict.
- Analysis that lives in one spreadsheet. The monthly pack is rebuilt by hand by one analyst. It is accurate, slow, and unavailable when they are on leave.
- Dashboards nobody opens. A wall of charts answers no particular question, so people go back to asking the analyst directly.
CAPABILITIES AND DELIVERABLES
What is involved,
and what you get.
RELEVANT CAPABILITIES
- Data foundations. Bringing source data together on a schedule, with the pipeline tested and failures visible.
- Metric definitions. Writing down how each measure is calculated and which source is authoritative, then holding to it.
- Semantic layer. A shared modelling layer so every tool and report calculates the same metric the same way.
- Reporting and dashboards. Dashboards built around the decisions people actually make, with freshness shown alongside the figures.
- Reconciliation support. Tracing where two numbers diverge and fixing the cause rather than adjusting the output.
- Decision-support tools. Small focused applications for forecasting, allocation or scenario comparison when a dashboard is not enough.
TYPICAL DELIVERABLES
- Data source inventory
- Metric definition catalog
- Modelled data warehouse
- Semantic modelling layer
- Decision-focused dashboards
- Data quality monitoring
WHAT COULD IMPROVE
- Meetings that debate the decision rather than the figures
- Monthly reporting that does not depend on one analyst's time
- Visible data freshness, so staleness is noticed before it misleads
- A model your analysts can extend without starting again
Possible improvements, not guaranteed outcomes.
EXAMPLE WORKFLOW
How an engagement
usually runs.
A typical shape. The order changes with the situation, and we say so when it does.
Start from the question
We collect the decisions people are trying to make, because that determines what is worth modelling.
Agree the definitions
We get the business to settle what each metric means and record it where everyone can see it.
Build the foundation
We bring the sources together, model them, and test the pipeline so freshness and failures are known.
Reconcile
We compare the new figures against the numbers people trust today and explain every difference.
Publish and enable
We release the dashboards, train the people who will use them, and hand over how to extend the model.
RELATED PATHWAYS
Where this service
usually fits.
Simplify Operations
Each request has a defined path, a visible status and a named owner. Routine steps run on their own, and exceptions reach the right person with the context attached.
Connect Business Systems
Priority flows move automatically between systems with a defined owner for each record, and failures raise an alert instead of passing unnoticed.
QUESTIONS
Questions we are
asked about this.
Not always, and it depends on how many sources you have and how often figures must be combined. If reporting comes from one system and nobody disputes the numbers, better use of that system's reporting may be enough. A warehouse starts to pay for itself once several sources must be joined, reconciled and kept historically.
Not as a separate project, because clean-up with no destination tends to drift. We would rather pick the decisions that matter most and clean the data those depend on, in that order. The messy areas nobody is reporting on can wait, and some can wait indefinitely.
We can usually tell you why they differ, which is more useful. Differences normally come from different filters, different dates or a different idea of what counts, and once that is visible the business can choose the definition it wants. The fix is then to apply that definition in one place rather than in each report.
Yes, in most cases, and that is usually our first suggestion. Most mainstream tools are capable enough, and the real problems are more often in the model and definitions underneath. If your tool genuinely cannot do what you need, we will explain exactly what it is blocking before suggesting anything else.
A CLEARER VIEW
Your next move starts with a clearer view.
Tell us what you want to improve, what is getting in the way, or what you want to build. We will help you explore a practical path forward.